Objects Color Segmentation and RSTC-Invariant Representation using Adaptive Color KLT and Modified Mellin-Fourier Transform

Roumen Kountchev and Roumiana Kountcheva


Color object segmentation, Adaptive color KLT, Modified 2D Discrete Mellin-Fourier Transform, RSTC-invariant object description


In this paper is presented one new approach for invariant object description, based on color segmentation with hyperboloid color model in the Karhunen-Loeve Transform space. The object description is represented as a vector calculated using a Modified 2D Discrete Mellin-Fourier Transform. The description of objects, extracted in result of the segmentation, is invariant to 2D rotation (R), translation (T), scale (S) and contrast (C) changes. In the paper are given some experimental results, which confirm the efficiency and the reliability of the new approach and the abilities for its application in image retrieval systems. The main advantages of the new approach are outlined, based on comparison with other similar methods.

Important Links:

Go Back